microsoft/onnxruntime

Pull Requests Commits

Address review comments for WebGPU QuantizeLinear/DequantizeLinear

edgchen1 committed 29 days ago

32d883ab

Merge remote-tracking branch 'origin/main' into edgchen1/webgpu_quantizelinear

edgchen1 committed 29 days ago

ab84e61e

use helpers in DQ, some refinements

edgchen1 committed 29 days ago

f9221d47

[QNN-EP] Add additional guards for file mapping (#27871)

quic-calvnguy committed 29 days ago

Verified a2742d0f

fix a security issue in SVM* (#27950)

xadupre committed 29 days ago

Verified ce92643c

Bump lodash from 4.17.23 to 4.18.1 in /js/web (#27990)

dependabot[bot] committed 29 days ago

Verified 3471a5cf

[WebGPU EP] Fuse QMoE 1-token decode path to reduce GPU dispatches (#27998)

qjia7 committed 29 days ago

Verified c36c4225

centralise feed authentication for ADO pipelines (#27997)

eserscor committed 29 days ago

Verified 87b06439

[VitisAI] fix dangling pointer (#27949)

amd-genmingz committed 29 days ago

Verified 5e555442

Fix plugin EP profiling timestamp skew on macOS (#27994)

tianleiwu committed 29 days ago

Verified bbb0cd0d

Add CUDA Graph support for the CUDA plugin EP (#28002)

tianleiwu committed 29 days ago

Verified 58a87dc1

address comments

edgchen1 committed 30 days ago

e3192308

Modify scale & offset of WhereDummyDq (#27109)

qti-hungjuiw committed 30 days ago

Verified ce91376b

[Plugin EP] Port graph capture/replay APIs (#27958)

adrianlizarraga committed 30 days ago

Verified 7afe4c25

ICM fixes (3/n) (#27925)

hariharans29 committed 30 days ago

Verified 3fad2939

WebGPU QuantizeLinear: add int4/uint4 support

edgchen1 committed 30 days ago

d9551413

mlas/arm64: add BF16 fast-math conv kernels for NCHW/NCHWc paths (#27878)

milpuz01 committed 30 days ago

Verified efc89b5f

webgpu support for qwen3.5 (#27996)

guschmue committed 30 days ago

Verified 3ac60400

Add pre-commit git hook to run lintrunner on staged files (#28013)

vraspar committed 30 days ago

Verified cd205243

Refactor QuantizeLinear shader to use standard helpers

edgchen1 committed 30 days ago

d3ae227a

Add bounds validation for LinearClassifier coefficients (#27989)

vraspar committed 30 days ago

Verified 9c90e791

Validate g_idx values in MatMulNBits to prevent OOB read (#27582)

vraspar committed 30 days ago

Verified 127704c4

WebGPU QuantizeLinear: per-axis, blocked quantization, Option B packing

edgchen1 committed 30 days ago

95f515a9

[CPU/CUDA ep] Improve DeformConv op performance (#27824)

ShirasawaSama committed 30 days ago

Verified 9d7e6d53

[webgpu] Set `is_channels_last` to true by default in `ComputeMatMul` (#27674)

Jiawei-Shao committed 30 days ago

Verified 9e3614b5

fix target_ids out of boundary in TreeEnsemble* (#27951)

xadupre committed 30 days ago

Verified c6ff87f6

WebGPU QuantizeLinear: add per-axis support and int8 fixes

edgchen1 committed 31 days ago

65cf878e

Remove unnecessary model package test (#28015)

chilo-ms committed 31 days ago

Verified c1596034

edgchen1 committed 31 days ago

39bf86a2

Model Package Support (#27786)

chilo-ms committed 31 days ago

Verified fbba40a4

Older